The error looks like this now:
It seems to be reasonably common that this happens because a non-parquet file is being queried in a parquet table.
The error message should say something like "File 'hdfs://nn/books/books1G.csv' has an invalid Parquet version number: .99 and does not appear to be a valid Parquet file. This could be due to stale metadata. Try running "refresh s3db.books_s3"